Table 5.
Between-arm differences on WEMWBS at 12-months post-randomisation with and without accounting for adherence or attendance.
| Model (47 family carers within 39 families) | Adjusted mean difference* | Lower 95% CI | Upper 95% CI |
|---|---|---|---|
| Two-level model | 3.96 | −1.39 | 9.32 |
| Single-level model† | 4.38 | −1.02 | 9.78 |
| IV regression accounting for adherence† | 5.05 | −0.70 | 10.79 |
| Maximum efficacy based on family session attendance | 5.84 | −0.80 | 12.40 |
| Maximum efficacy based on main family carer attendance | 6.84 | −0.84 | 14.53 |
| Maximum efficacy based on main family carer attendance, main family carer responses only (n = 36) | 4.85 | −3.71 | 13.41 |
Adjusted for baseline WEMWBS score and site.
Cluster robust standard errors account for clustering of participants within families.